Skip to main content

Table 1 Input parameters to the model

From: Modeling of smartphones’ power using neural networks

Predictor name

Description

Generation method

Range

MF utilization

CPU utilization while operating at the median frequency rang

Variable polling with OS counters

0–1

HF utilization

CPU utilization while operating at the high frequency range

Variable polling with OS counters

0–1

Screen on

Fraction of time screen was on

Broadcast receivers

0–1

Screen brightness

Average screen brightness

Variable polling

20–255

Call ringing

Fraction of time smartphone was ringing

Broadcast receivers

0–1

Call off-hook

Fraction of time smartphone was in call

Broadcast receivers

0–1

Data on

Fraction of time smartphone was connected to some mobile network

Broadcast receivers

0–1

Data traffic

Average number of bytes sent/received through mobile network per second

Variable polling with OS counters

≥0

WIFI on

Fraction of time phone is connected to some WIFI network

Broadcast receivers

0–1

WIFI traffic

Average number of bytes sent/received through WIFI interface per second

Variable polling with OS counters

≥0

SD traffic

Average number of sectors read/written per second

Variable polling with OS counters

≥0

Audio on

Fraction of time audio device was active

Broadcast receivers

0–1

GSMRSSI

Average mobile received signal strength indication

Broadcast receivers

(−113)–(−48) dBm

NET HSDPA

Fraction of time mobile connected to HSDPA network (3G)

Broadcast receivers

0–1

NET EDGE

Fraction of time mobile connected to EDGE network (2.5G)

Broadcast receivers

0–1

NETGPRS

Fraction of time mobile connected to GPRS network (2G)

Broadcast receivers

0–1

GPS on

Fraction of time GPS adapter is on

Broadcast receivers

0–1